Readnihongo app icon

ReadNihongo

Stories you can finish

ReadNihongo / The learning design

Yes, we read the footnotes.

The science behind your reading

A little obsessive.
About your Japanese.

How much of the Japanese you read becomes Japanese you remember? That question is behind the stories, the word lookups, and the reviews that bring you back.

We dig into the research on reading, vocabulary, and memory, then sweat the details of putting it into practice. You get to enjoy the story.

The learner reveals a word meaning while reading.
One detail worth obsessing over01 / Word support

A small word meaning.
A measurable difference.

Mean target words learned (%)

With word meanings45.3%
Without word meanings26.6%

+18.7percentage points
with glossed reading

Yanagisawa et al. (2020) · 42 studies · 3,802 learners. Pooled target-vocabulary means; delayed intervals varied. The two test pools are not a single forgetting curve. These are study results, not ReadNihongo measurements.
73

studies in a recent
extensive-reading review

R01
3,802

learners in the review
of vocabulary glosses

R07
3,411

learners in the review
of spaced practice

R06

How a reading study works.

Follow an illustrated example of how researchers test whether word meanings help learners remember vocabulary.

Researchers prepare two versions of a passage and add short word meanings to one version.

01 / Start with the evidence

Reading does real work.
Across more than one skill.

A 2025 meta-analysis found positive short-term effects across seven outcome areas. The scale below shows the size and uncertainty of each estimate.

The learner turns to the next page of a story.

Extensive reading versus comparison instruction

Sangers et al. · 2025
OutcomeCohen’s d · 95% confidence intervalEstimate

Reading comprehension52 comparisons

Extensive-reading outcomes: Reading comprehension0.31

Vocabulary31 comparisons

Extensive-reading outcomes: Vocabulary0.50

Decoding & fluency23 comparisons

Extensive-reading outcomes: Decoding & fluency0.38

Motivation17 comparisons

Extensive-reading outcomes: Motivation0.34

Writing14 comparisons

Extensive-reading outcomes: Writing0.31

Oral proficiency11 comparisons

Extensive-reading outcomes: Oral proficiency0.42

General proficiency9 comparisons

Extensive-reading outcomes: General proficiency0.53
Inspect the plotted numbers
Extensive-reading outcomes: Cohen’s d and reported 95% confidence intervals
OutcomeEstimate95% CIComparisons
Reading comprehension0.310.16 to 0.4752
Vocabulary0.500.24 to 0.7731
Decoding & fluency0.380.1 to 0.6623
Motivation0.340.004 to 0.6817
Writing0.310.07 to 0.5614
Oral proficiency0.420.21 to 0.6411
General proficiency0.530.22 to 0.849
Table 1, p. 15. Points are pooled effects; lines are reported 95% confidence intervals. Effect sizes use standard-deviation units, rather than percentage gains. Comparisons overlap across outcomes. “Oral proficiency” combines listening, speaking, and grammar measures. The review primarily concerns English learning, with two Spanish interventions.
Our call for ReadNihongo

Keep Japanese stories at the center. Offer level choices so you can find a readable starting point, and put word meanings within reach when a sentence gets in the way.

02 / Make it available later

The moment you try
to remember matters.

In a vocabulary experiment, students learned 40 Swahili–English pairs. After their first successful recall, researchers varied whether they kept studying, kept retrieving, or stopped both.

One week later, the groups that kept retrieving recalled about 80% of the pairs. The other groups recalled 36% and 33%.

R55
The learner pauses to remember a word and then checks the answer.
Our call for ReadNihongo

Let you save a word from a story, then come back and try to recall it. In review, give yourself a moment before checking the choices or revealing the answer.

The test was one week later02 / Recall

What survived the week?

Word pairs recalled (%)

Keep studying + keep recalling≈80%
Stop studying + keep recalling≈80%
Keep studying + stop recalling36%
Stop both after first success33%
Karpicke & Roediger (2008), p. 967. Approximate means as described in the paper; error bars are not reconstructed. This word-pair experiment motivates recall practice, rather than predicting a Japanese-reading gain.

The timing gets attention, too

An encounter.
A gap.
Another chance.

A bird delivers a vocabulary card for another review.

Spaced practice has support across 48 experiments involving 3,411 learners. A separate review found a positive relationship between repeated encounters and vocabulary learning across 26 studies and 1,918 learners.

So we give saved words a way back into your day. Review intervals respond to your answers: forgetting a word brings it back sooner; successful recall gives it more room.

R06 R05
10 minreturn interval after “Again”
≥1 dayminimum after “Good”

These are settings in ReadNihongo’s current review scheduler. The research supports spacing; it does not establish these exact intervals as optimal.

03 / The small decisions count

Yes, we read
the null results, too.

A review of 30 studies found that the comprehension benefit of reading with audio depended on pacing. In the self-paced subgroup, the confidence interval crossed zero.

The learner starts listening and follows along with the text.

Adding audio: the pacing detail

Clinton-Lisell · 2023
OutcomeHedges’ g · 95% confidence intervalEstimate

Experimenter-paced

Audio-assisted versus reading-only comprehension: Experimenter-paced0.41

Self-paced

Audio-assisted versus reading-only comprehension: Self-paced0.06
Inspect the plotted numbers
Audio-assisted versus reading-only comprehension: Hedges’ g and reported 95% confidence intervals
OutcomeEstimate95% CI
Experimenter-paced0.410.13 to 0.7
Self-paced0.06-0.07 to 0.19
Abstract, p. 133. Subgroup estimates and reported 95% CIs; mixed first- and second-language studies. The self-paced result does not establish a reliable comprehension advantage. This outcome does not measure every possible use of audio.

Useful tools.
With a reason to exist.

Audio, when you want it.

Hear a Japanese sentence, replay a sound, or read at your own pace. You choose when listening helps.

Furigana, when you need it.

Get a pronunciation cue, then try the passage without it when you are ready. The goal is growing independence.

English, within reach.

Resolve the meaning, then return to the Japanese. Keep the story moving.

R19 R41 R42 R47

And yes, Japanese is its own problem

Kana. Kanji. Context.
The details change.

A sliding panel pairs 山 with やま and a drawing of a mountain.

We also look at research on Japanese readers themselves. Mohar’s beginner study followed 13 learners over 18 weeks and 190 texts, exploring which materials and reading targets were workable. Other Japanese studies examine graded texts, script support, and transfer to new passages.

That is why level choice, optional furigana, sentence audio, and accessible meanings all have a place here. A result from English learning can inform a principle; Japanese learners still deserve Japanese-specific attention.

R28 R30 R33 R36

The question behind the features

Will this help
your Japanese travel?

From this story to the next one. From a supported sentence to something you can follow on your own.

The learner asks a question and the scientist explains.
What are you optimizing for?

Japanese you can follow today, remember later, and recognize in an unfamiliar passage. That means balancing reading flow, useful assistance, and later recall. The figures on this page are published study findings that inform those choices. Evaluating the complete ReadNihongo experience requires its own learner studies.

Does reading take care of speaking and listening, too?

Reading builds a valuable foundation, and broader language benefits appear in some studies. Conversation and listening still deserve their own practice. We care about which skill was actually tested, and about understanding new Japanese as well as getting comfortable with a familiar story.

R28 R60
What about furigana and English translations?

They help you reach Japanese you might otherwise abandon. Furigana supplies a pronunciation cue; English can resolve a meaning. Use the help, then try returning to the Japanese with less of it. We keep these supports optional because the useful amount changes with the reader and the passage.

R19 R41 R42

04 / For our fellow footnote readers

Of course
we brought the sources.

The scientist checks and organizes the research papers.

All 80 references, grouped so you can follow a question deeper. The highlighted chart values were checked against the original papers. The wider reading list includes reviews, experiments, theory, theses, and official resources.

Research informs the design; the charts report study outcomes rather than ReadNihongo user results. Evidence varies by language, learner, comparison, and test. Reading notes updated .

Extensive reading and language development12 references
  1. R01
  2. R02
  3. R03
  4. R58
  5. R59
  6. R60
  7. R61
  8. R62
  9. R75
  10. R77
  11. R78
  12. R80
Vocabulary, repetition, and context16 references
  1. R04
  2. R05
  3. R13
  4. R14
  5. R15
  6. R16
  7. R17
  8. R18
  9. R43
  10. R52
  11. R57
  12. R69
  13. R72
  14. R73
  15. R74
  16. R76
Spacing, retrieval, and memory6 references
  1. R06
  2. R53
  3. R54
  4. R55
  5. R56
  6. R70
Glosses, translations, and text difficulty13 references
  1. R07
  2. R08
  3. R09
  4. R10
  5. R11
  6. R12
  7. R19
  8. R20
  9. R21
  10. R22
  11. R23
  12. R24
  13. R71
Audio and reading fluency8 references
  1. R25
  2. R26
  3. R27
  4. R28
  5. R29
  6. R37
  7. R47
  8. R79
Japanese reading, kanji, and learner differences19 references
  1. R30
  2. R31
  3. R32
  4. R33
  5. R34
  6. R35
  7. R36
  8. R38
  9. R39
  10. R40
  11. R41
  12. R42
  13. R44
  14. R45
  15. R46
  16. R48
  17. R63
  18. R64
  19. R65
Grammar and attention3 references
  1. R49
  2. R50
  3. R51
Reading on screens and on paper3 references
  1. R66
  2. R67
  3. R68

Found a paper we should read, or a claim worth revisiting? Send it our way. Our editorial standards.

That is the thinking. Here is the fun part.

Find a story.
Let the Japanese begin.

Something interesting to read. Help when you need it. A reason to come back.

Start reading
The learner chooses another story and starts reading.